Development discussion for DB Browser for SQLite (previously SQLite Browser)
Looking at the alpha1 release notes, #1784 seems like it was given only a cursory mention instead of a proper entry. eg we mostly missed it. :frowning:
Let's make sure it's included with a proper full line entry in the final release notes. :smile:
Let's make sure it's included with a proper full line entry in the final release notes. :smile:
:thumbsup:
Everytime I think of "releases", head goes back to one of the original (early 90's) email jokes that circulated. http://gradha.sdf-eu.org/textos/klingon_programmer.en.html #6 seems applicable. :D
Everytime I think of "releases", head goes back to one of the original (early 90's) email jokes that circulated. http://gradha.sdf-eu.org/textos/klingon_programmer.en.html #6 seems applicable. :D
Heh Heh Heh
This machine is GAGH! I need dual Pentium processors if I am to do battle with this code!
Sitting next to me is an old dual Nocona (Xeon) machine. Technically it's functional, though I've taken out the power supply and used it in another box. But if power supply is put back in, it runs.
It runs Win7 reasonably ok, with it's ECC 5GB of ram. Linux doesn't boot on it, though FreeNAS does.
It was probably made in about the same era as that article. :)
Well, not quite. Nocona Xeon's were from ~2004. But similar level of ancient feel.
case in point: https://www.sqlite.org/syntaxdiagrams.html
@scottfurry Yes, antlr2 is no longer a requirement. I replaced the antlr-based parser by a bison-based parser. So we no longer need the antlr2 library. Also the new bison parser doesn't depend on any extra library which means we don't have to add a new dependency instead.
You only need bison when changing the parser itself in order to regenerate the C++ files.
That was one of the intended benefits of the new parser by the way.
No major bugs have been reported for the new DB4S release for the first day, so I just updated the
currentrelease file in our GitHub repo to point at it. That'll kick off the "new version is available" message for people running 3.10.x and older. If things still look good tomorrow, I'll do the same for the currentreleasefile on our download server. That'll show the notification for anyone running 3.11.x onwards. :smile:
Meh, just updated the one on our download server now instead. Download server doesn't seem to be having any issues, though I should probably read the
nload man page and figure out how to change the graph limits. eg this is the download server a few seconds ago (inbound and outbound):
@justinclift @MKleusberg - I got the Cherry-Pick instructions. However, this raises a question in my mind. Does this not lead to multiple PR's? There would be the PR to the master branch. Once that gets approved, would another PR be needed to merge something into the 3.12.x branch where situation warrants?
Not generally. With most projects, the PR is done to the head of the main development branch.
master in our case. That ensures the change is baked in, and part of all future releases. With that done, one of project committers can cherry-pick it from the master branch to whichever other branches need it.
With our project, we seem to really only do stuff with the last major release branch (eg
v3.12.x now) and master, so it's fairly straight forward. :smile:
Probably need to modify Github Issue template: '3.12.0-(alpha or betas)'
Ok, the PortableApp for 3.12.0 is online now too. I think all of the bits needed for it are done.
@justinclift @MKleusberg @chrisjlocke https://www.kdab.com/fixing-a-common-antipattern-when-loading-translations-in-qt/ Seems we can do something different to respect user's locale WRT QTranslator in src/Application.cpp
Heads up. Qt Korean group has published an English/Korean Qt PDF book - http://qt-dev.com/board.php?board=kkknotice&command=body&no=48&
Took a look, but there's religious brainwashing/indoctrination crap all through it. eg "Jesus loves you" stuck prominently in the header on every page, quotes from the bible embedded, etc.
That's a real shame, as it could otherwise have been good content for people. In it's current form, very much avoid.
That's a real shame, as it could otherwise have been good content for people. In it's current form, very much avoid.
@justinclift Re: #2321 - this one seems a touch vexing. I just looked at the DB4S dependencies on Linux and qt5svg doesn't show. But I recall we went back and forth about svg images (mostly for desktop purposes). I suspect the problem encountered is something to do with the default image handlers qt (or us for that matter) that are loaded. And a missing handler for svg images caused the crash (dll not packaged). That's my theory, anyway.
@scottfurry For the crash aspect, I'm thinking it was more like this:
- Qt loads
plugins/imageformats/qsvg.dllfrom the DB4S install directory. Our Qt dll's at the moment are version 5.12.8. Instead of thatqsvg.dllbeing all that's needed, that dll turned out to have a dependency onQt5Svgd.dll(which we weren't packaging). - So, Qt asks the OS (or whatever) to find and load that
Qt5Svgd.dll. - For the case where the user already has an install of Qt 5.9.2, the
Qt5Svgd.dllfrom that was being picked up and loaded. - The crash then happens when DB4S tries to use SVG, as Qt 5.9.2 libraries aren't compatible with Qt 5.12.x ones.
The fix was simply to include the correct version of
Qt5Svgd.dll with our packaging, so there's no need for the OS (or whatever) to go looking around for a potentially different one.
If that doesn't work either, let us know. :)